At Northeastern’s Network Science Institute, I am engineering an end-to-end HPC pipeline (SLURM) over 500GB+ of USPTO patent data, handling ETL, normalization, and entity resolution with Pandas, DuckDB, and lxml. I train Doc2Vec and neural encoders (Gensim, PyTorch) to embed patents, then build dynamic inventor–patent–CPC networks and compute community detection, assortativity, and centrality trajectories using NetworkX. I also apply causal inference (event-study, DiD) to quantify how shifts in intellectual property regimes reshape collaboration intensity and cross-disciplinary ties.